Noor Yadallee
Innovation enthusiast
Quatre Bornes, Mauritius
Actions
Has-been frog in a well, on a quest to broaden his horizons, face challenges head on and hoard new experiences.
Links
Area of Expertise
JAX & Flax: From Fundamentals to Recreating Research
This talk is for anyone interested in building and understanding modern machine learning models. We'll start with a quick introduction to JAX, highlighting its core principles like automatic differentiation and JIT compilation, which make it so powerful for high-performance computing.
We'll then dive into Flax, a neural network library built on top of JAX. You'll learn how Flax simplifies model building with its flexible module system. Instead of just talking about it, we'll get hands-on.
The second half of the session is a live coding tutorial where we'll recreate a research paper's model from scratch. Attendees will follow along, gaining practical experience in using JAX and Flax to implement a real-world architecture. By the end, you'll not only understand the "what" and "why" behind JAX and Flax but also have the skills to start building and replicating your own research models.
We need to talk about Google's Jax
JAX has emerged as a powerful and increasingly essential tool in the world of machine learning, particularly for high-performance numerical computation and large-scale model training. This session dives deep into JAX, exploring its core functionalities and showcasing its potential to revolutionize your ML workflows.
We'll start by understanding what makes JAX unique: its functional programming paradigm, automatic differentiation capabilities (both forward and reverse), and the ability to seamlessly target various hardware accelerators (GPUs, TPUs). We'll unpack how these features combine to enable efficient and scalable computation.
Beyond the fundamentals, we'll delve into practical applications. This session will demonstrate how JAX empowers researchers and practitioners to build and train complex models with ease. We'll explore real-world use cases, including an examination of how JAX plays a crucial role in powering cutting-edge solutions like Google's Gemini.
This session is designed for those with a foundational understanding of machine learning and some familiarity with Python. While prior experience with JAX isn't required, a willingness to explore new concepts and a passion for pushing the boundaries of ML are essential. Join us to discover why "We Need to Talk About Google's JAX" and how it can unlock the next generation of your AI projects.
The state of Computer Vision on GCP
Leveraging Computer Vision on GCP from using pre-built ML models to building custom image classifiers using linear models, deep neural network (DNN) models or convolutional neural network (CNN) models.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top